def __init__(self, exp_desc):
"""
:param exp_desc: an experiment descriptor object
"""
assert isinstance(exp_desc, ed.ExperimentDescriptor)
self.exp_desc = exp_desc
self.exp_dir = os.path.join(misc.get_root(), 'experiments', exp_desc.get_dir())
self.sim = misc.get_simulator(exp_desc.sim)
import numpy as np
import matplotlib
import matplotlib.pyplot as plt
import misc
import inference.mcmc as mcmc
import inference.diagnostics.two_sample as two_sample
import simulators.gaussian as sim
import experiment_descriptor as ed
import util.io
import util.math
import util.plot
root = misc.get_root()
rng = np.random.RandomState(42)
prior = sim.Prior()
model = sim.Model()
true_ps, obs_xs = sim.get_ground_truth()
# for mcmc
thin = 10
n_mcmc_samples = 5000
burnin = 100
def get_true_samples(seed):
"""
Generates MCMC samples from the true posterior.
import os
import numpy as np
import ml.models.neural_nets as nn
import ml.trainers as trainers
import ml.loss_functions as lf
import simulators.lotka_volterra as sim
import misc
import util.io
import util.plot
dir = os.path.join(misc.get_root(), 'results', 'lotka_volterra', 'other',
'failed_sims_model')
def gen_data(n_data=100000, rng=np.random):
"""
Generates training data to fit the model.
:param n_data: number of datapoints
:param rng: random number generator
"""
res_file = os.path.join(dir, 'data')
if os.path.exists(res_file + '.pkl'):
ps, ys = util.io.load(res_file)
else:
